Comment

1 September 2005

Informed decision

Users need to marry correct technology to application.

By Garrett Schmidt

As wireless technology matures in the industrial world, the number of transmission protocols and network topologies continues to proliferate. Where wireless was once a point-to-point cable replacement device, the capabilities of wireless devices are growing in terms of speed, distance, cost, transmission method, and networking capabilities. End device connectivity now ranges from network component to sensor level.

Choosing a wireless device is often in the interest of one thing: saving money. Also eliminating cable and conduit represents an astronomical savings in time. Digging, drilling, backfilling, and repaving are no longer the lion's share of a budget. Choosing a wireless device could mean more funding can go toward higher quality system components and a faster deployment of those components.

Wireless devices see use in a wide array of applications, and so it only makes sense there is a wide array of different wireless technologies geared toward those applications. It is here that users find difficulty when they have to pair the correct technology with the appropriate application.

For instance, infrared red (IR) devices are well suited to remote control applications in your living room, but it a poor choice in a dusty, obstructed factory environment. A radio frequency (RF) device is a much better choice there because you have to overcome dust and obstructions. So what RF technology is correct for an application? It becomes much easier to select a technology if the industrial application space is broken into several segments. There are clear definitions of the devices found in each level, and so are the wireless technologies. The requirements of a wireless device for networking PCs are obviously not for connecting a sensor to a PLC. All of these devices operate utilizing license-free spread spectrum technology, meaning they operate within a set of parameters defined by the FCC.

License restrictions

Operation of a wireless device without a license imposes several strict regulations. The device must operate within the bands 902-928MHz, 2.4-2.4835GHz, or 5.725-5.850GHz (although users don't often use this band for industrial applications). The transmit power of these devices must not exceed 1 watt, and the gain of an antenna cannot exceed 6dBi. Part 15.247 of the FCC regulations also requires the use of one of two possible modulation techniques: Frequency Hopping Spread Spectrum (FHSS) or Direct Sequence Spread Spectrum (DSSS).

Frequency hopping radios function by constantly changing the transmit frequency, moving in a pseudorandom pattern around the frequency spectrum at a high rate of speed. Therefore, the actual bandwidth usage during a transmission is very narrow, usually only a few dozen kilohertz, so the RF power is very concentrated. This transmission method has a high interference tolerance since the power spreads across the spectrum and not the user data, thus guaranteeing that information will pass between the radios until 100% of the frequency band clogs with RF noise. Frequency hopping is also inherently secure; the moving target is changing anywhere from 50-100 times per second, and is extremely difficult to track and lock onto. Typically, users who employ this technique require a relatively low baud rate, which allows for a great amount of energy per data bit improving the probability of successful radio transmissions. There are proprietary radio systems in the industry that implement FHSS to move serial data or I/O updates over the air. Bluetooth devices (802.15.1) also use frequency hopping.

DSSS is another transmission method employed in the ISM bands. The main use for DSSS is moving large packets of information, but it is more prone to interference since it operates across a wider signal channel. Direct sequencing starts with the transmitter replacing every bit of user data with a random code word. The code word is de-spread, leaving the original bit intact. Interference that enters the packet suppresses at the receiver when the code word is de-spread, but a certain jamming margin exists. The DSSS radio will function perfectly until it exceeds the jamming margin, at which point the radio will shut down completely. For this reason, DSSS is better suited for low to medium interference environments, 100% of the data will pass, and larger packet sizes can go out at higher baud rates. Wireless Ethernet (802.11) devices, as well as ZigBee devices (802.15.4), employ direct sequencing.

At the plant level, the wireless device must satisfy the LAN protocol and be capable of passing massive amounts of network data at high speeds of 10 or 100 Mbps. The network will probably consist of a combination of hardwire with wireless devices installed where cabling is too difficult to run, and connections between the plant and area level will likely also utilize the same type of wireless device for ease of integration. Integrity is critical at these application levels, as a radio failure or data error will bring down the entire network. As such, an IT department will be heavily involved with the integration of the radio onto the network, and issues of security, accessibility, and "known" technology will undoubtedly arise. The types of wireless devices for these application levels are very limited, mostly due to speed.

Standard deviations

The variations of IEEE standard 802.11 are the most accepted and well-known types of wireless devices on the market today. In fact, many find the word 'wireless' synonymous with 802.11a, b, g, or the recently ratified 802.11i. In an industrial application, the plant/area level application is the best fit for an 802.11 device. 802.11b or g are the most common variations of the standard; they operate in the 2.4GHz band at data rates up to 54 Mbps for distances up to several hundred feet. 802.11a also has data rates up to 54Mbps but operates in the 5.8GHz band. The advantage of these is they are relatively inexpensive, and devices from different manufacturers are interoperable, but this is also a drawback because they are not very secure. They utilize Wired Equivalent Privacy (WEP), a security feature with many flaws, one of which is any amateur hacker can defeat it. 802.11i has recently come onboard and addresses those flaws in previous versions of 802.11 with Advanced Encryption System (AES). It also features Wi-Fi Protected Access (WPA), which replaces WEP to enable backwards compatibility with the older 802.11 versions.

Progressing into the cell level, a wireless device typically will not need to run at network speed, and transmission distances may be longer as devices become more distributed. The user is still transporting network data, but the quantity of data is much less, and the cell level protocol will be different than the plant level. The network may be Ethernet or serial; errors may or may not recover depending on how critical the application is, but a failure will shut down a large portion of the overall network. Many more choices for a wireless device exist for this application level, both as standard and proprietary over-the-air protocols for connecting the cell devices to the field devices.

The user has the greatest flexibility in choosing a wireless device for this application level. Within this sector, you can choose a wireless device to fit an application perfectly in terms of transmission distance, over-the-air data rates, physical connections, and protocol support. A wireless device can become a network device or just a cable replacement. Serial and Ethernet connective devices are available on the market, and several products even combine the two, operating as a wireless terminal server.

RF technology options differ greatly in the cell/field layer. Several standards exist that may fit the application well, but most wireless devices use a proprietary transmission scheme. A manufacturer may use 802.11 in a wireless Ethernet function, but it typically isn't well suited because the network devices are generally scattered beyond the range of an 802.11 device (the standard indicates 100mW transmit power) unless external amplifier and high gain antennas are used. A proprietary Ethernet solution is a better choice, since the allowable transmit power is 1 watt and the over-the-air rate is less than 1 Mbps, increasing both the transmit distance capability and reliability of the wireless connection.

Since most serial RF modems are proprietary, Bluetooth (IEEE 802.15.1) is the only applicable wireless standard. Bluetooth comes in commercial and industrial applications and works well in both. In an industrial context, it transmits RS-232 and RS-485 protocols at nearly 1Mbps. The only limitation lies in the effective range, which is in hundreds of feet for a Class 1 Bluetooth device at 100mW, or in tens of feet for a Class 2 Bluetooth device at 10mW. It has a very low latency, making it a good choice for high-speed serial bus protocols and controllers.

Proprietary serial radios are available in many forms, ranging from OEM boards to desktop boxes to DIN rail mount products. Their features and functions have just as much variety; simple versions simply eliminate the cable and distance limitations of RS-232 and RS-485, while others actually eliminate the need for a field device by providing direct connections for sensors and are addressable via the field level protocol. The majority of these devices are frequency hopping in the 900 MHz band at 1 watt of transmit power, making them ideal for sending serial protocols a few miles over the air.

Connecting field devices to the element or sensor level is an area currently targeted by the manufacturers of wireless devices. Because the sensor is only providing a few bits or bytes of information, the transmission speed of a wireless sensor device is much lower than any other part of the network. Errors are not critical because in any given time frame (in terms of wireless latency, measured in milliseconds) the measured value cannot change drastically enough to cause an alarm state (with the exception of pulse counting). The next update will replace the incorrect value, while a failure of the wireless device only affects the connected sensor and should trigger a local alarm.

RF technology on the sensor level is not yet well developed. Wireless standards that seem to fit include Zigbee (IEEE Standard 802.15.4) and Bluetooth, which are available on a single chip in order to reach the number one goal of sensor radio developers: ultra low cost. The Zigbee standard won ratification in early 2005 and is most noticeable for its low power consumption, which lends itself well to battery powered devices. However, the throughput is very low as a battery-powered device cannot be available for continuous polling, it will spend most of the time in sleep mode, awakening perhaps only once a day for an update transmission. Bluetooth sensor devices are not available on the market, the higher power consumption means they won't last long on a battery. Several proprietary high power sensor devices are on the market, most of which operate on 900MHz at the FCC imposed limit of 1 watt. These radios are not battery operated and provide continuous updates for 4-20mA analog and digital signals in different multiplexer configurations. They send signals thousands of feet or more and obviously are not low cost solutions, but are robust, industrially hardened, reliable re-placements for sensor wires.

Many different types of spread spectrum wireless devices can work throughout industrial application spaces, but there are some features that all of the devices should have in common to optimize their performance. The RF platform should contain several stages of filtering, the most basic being one that rejects all signals outside of the entire spectrum. The next stage will introduce adjustable in-band filtering. On a direct sequence radio platform, this filter will remove any interference outside of the specific operating channel; on a frequency hopping platform, the filter will remove any interference outside of the individual hop frequency. These filters need to be adjustable to account for frequency drifting due to component tolerances and ambient operating conditions. Likewise, due to temperature fluctuations, you should take oscillator drift into account.

The finished wireless device also needs to have a few key features in order to operate in an industrial environment. Power levels and availability vary from application to application, so it is important the device accepts a wide range DC supply voltage of 9-30 volts. This will ensure solar panels or power supplies can power the radio and operate in different industry segments that utilize lower supply voltages. The radio should also incorporate an RF Link relay contact to provide some deterministic reporting on the status of the radio communications. A relay contact would close some predetermined threshold of the radio signal strength to indicate a quality link between radios and open below the threshold to indicate the radios have lost communication. The contact can tie to a data logging device, controller, or visual or audible alarm to monitor the condition of the wireless network.

Wireless technology is rapidly gaining acceptance in industrial applications, and manufacturers are spawning new wireless device at an accelerating rate. With so many radio products currently on the market, choosing the correct one for an application is no easy task. Understanding the differences between data rates, distance, cost, transmission method, networking structure, and end device connectivity will ease the decision.W

Behind the byline

Garrett Schmidt is a product specialist at Harrisburg, Penn.-based Phoenix Contact Inc., working with products covering the industrial wireless products area. He has a Bachelor of Science degree in electrical engineering and is an ISA member.